1. Question 1

Forecasting CO₂ emissions using multiple input variables:

• ✅ Multiple regression

• ❌ Logistic regression

• ❌ Simple regression

• ❌ Polynomial regression

Explanation:
Multiple regression is used when predicting a continuous output using multiple independent variables.

---

2. Question 2

When is simple regression appropriate?

• ✅ Predicting annual rainfall based on average temperature.

• ❌ Customer segmentation

• ❌ Predicting sales with multiple factors

• ❌ Text classification

Explanation:
Simple regression works when there is one input variable and one continuous output.

---

3. Question 3

Predict employee productivity using multiple independence factors:

• ❌ Simple logarithmic regression

• ❌ Simple regression

• ✅ Multiple linear regression

• ❌ Simple polynomial regression

Explanation:
Multiple linear regression allows using several predictors at once.

---

4. Question 4

When is logarithmic regression useful?

• ❌ Monthly expenses with consistent growth

• ❌ Ice cream sales vs temperature

• ❌ Linear increase in sales

• ✅ Slow website traffic growth with added marketing budget

Explanation:
Logarithmic regression fits processes that grow quickly at first then slow down.

---

5. Question 5

Logistic regression shows high false positives. What can improve this?

• ❌ Collect more data

• ❌ Regularization

• ❌ Add more features

• ✅ Tune the classification threshold

Explanation:
Adjusting the threshold (e.g., from 0.5 to 0.6) changes the trade-off between false positives and false negatives.

---

6. Question 6

Logistic regression prediction = 0.65 means:

• ❌ 100% chance of return

• ❌ 0% chance of return

• ✅ 65% likelihood the customer will return the item

• ❌ 35% likelihood

Explanation:
Logistic regression outputs a probability between 0 and 1.

---

7. Question 7

Which scenario has highest log loss?

• ❌ Correct class = 0.9 probability

• ❌ Correct class = 0.7 probability

• ✅ Correct class = 0.1 probability (model is very wrong)

• ❌ 0.5 vs 0.5

Explanation:
Log loss heavily punishes being confident and wrong.
Predicting 0.1 for the correct class = maximum penalty.

---

🧾 Summary Table

---